Paul Sandoz wrote:
> Kohsuke Kawaguchi wrote:
>> Paul Sandoz wrote:
>>
>>> Note, I think there may be more mileage improving, if possible, the
>>> output buffering for UTF8XmlOutput. Having a private buffer and
>>> non-synchronized write method can significantly improve performance.
>>
>>
>> Yes, I should try this. In such a way that JVM can inline the method.
>>
>
> Yes. If you like i would be happy to do this.
Thanks. That would be great!
> Tomorrow i will also write a null XmlOutput implementation and measure
> the performance to see what the fixed cost of the JAXB serialization
> layer is.
Looking forward to seeing the data.
--
Kohsuke Kawaguchi
Sun Microsystems kohsuke.kawaguchi_at_sun.com